Skill UI

A comprehensive skill for capturing and documenting web application and graphical UI states during the development cycle. It supports full-page captures using Playwright, reliable post-processing cropping with PIL, capturing interactive states (hovers, selections), and even handles non-web desktop applications using mss/ctypes.

This skill captures and persists critical user decisions, project constraints, and architectural choices across multiple sessions and agents. It saves this context into a structured memory database (`mdbase`), ensuring that future development steps or agents can retrieve the original intent without re-deriving it from scratch. Ideal for documenting project rules, tech stacks, and design rationale.

This skill provides a direct, fast shell invocation to capture the current git diff (staged, working-tree, or specific file list) as raw text. It is designed to replace slower LLM agent steps that merely execute 'git diff' commands, significantly speeding up CI/CD workflows and meta-steps that require repository state analysis.

A meta-skill designed to monitor long-running processes, such as model fine-tunes or CI builds, executed within a detached tmux session. It captures output, automatically uses a sub-agent to diagnose failures, proposes a fix, and records the complete debugging process to memory. Ideal for intermittent failure debugging and continuous integration pipelines.

Designed to capture a fully functional, standalone static HTML file from any modern web application (React, Next.js, Vue, etc.). It resolves all computed styles and inlines CSS and images (as Base64) into the output, ensuring the captured view can be shared or archived without external dependencies. Supports both automated snapshotting via Puppeteer and interactive capture.

Provides comprehensive, programmatic control over tmux sessions. Users can list, capture output from, send specific keystrokes, and programmatically navigate panes/windows in remote, long-running interactive CLI sessions. Essential for monitoring AI code generation tools (like Claude/Codex) or managing complex multi-terminal workflows via scripting.

Atlas Ledger is a critical companion designed to provide lasting 'memory' to project contracts. It automatically processes detected deviations or 'drift' (e.g., from audits or reviews) and distills these instances into formal, actionable contract clauses (WHEN/DON'T/INSTEAD). These clauses are written to a persistent project memory file, Atlas.md. This mechanism ensures that lessons learned from failure are systematically captured, allowing future contract cycles to incorporate these rules, thereby strengthening overall project governance and reliability.

This skill guides you through a comprehensive, multi-phase interview to capture every necessary detail of a new project idea. It is designed to run BEFORE any code is written, forcing detailed specification gathering across purpose, features, data models, and technical constraints. The output is a clean, self-contained prompt.md file, ensuring subsequent development is based on solid, unambiguous requirements and preventing costly mistakes.

A powerful command-line tool for recording the macOS screen and capturing system audio simultaneously. Unlike QuickTime or standard `screencapture`, this tool captures raw display and system audio directly via ScreenCaptureKit, eliminating the need for complex virtual audio drivers (like BlackHole) or elevated permissions. It is designed for scripted content creation, automated demos, and complex recording scenarios where system audio is required via the CLI.

Use Mailtrap Email Sandbox to capture and inspect outbound emails without sending them to real recipients. This is crucial for development, staging, and CI/CD environments. It allows developers to test email content, headers, attachments, and spam scores using dedicated APIs, SMTP settings, or SDKs, ensuring mail integrity before deployment. Ideal for automation and rigorous quality checks.

A meta-skill designed to capture hard-won, non-obvious knowledge—such as complex debugging steps, multi-step operational workflows (e.g., deployment, migrations), or project-specific facts—and convert them into reusable Agent Skills. It ensures that future sessions, regardless of the user or team member, start with proven procedures, preventing the need to rediscover solutions from scratch.

This skill helps developers structure and record critical architectural decisions (ADRs) made during coding sessions. It automatically detects moments of decision, capturing the context, alternatives considered, and the rationale (trade-offs and consequences). This maintains a living log that explains 'why' the code base was built the way it is, preventing knowledge loss and improving project maintainability.